MSA 2050 disable cache flush

[Mike91]

Hello,

I want to turn off cleaning ssd cache when power is lost. As I understand it, this setting when executing the command: show cache-parameters is displayed as Cache Flush: Enabled. In the CLI reference guide, I did not find this option disabled.

thank advance

[arun_r]

Hi Mike,

I guess you are referring to SSD read cache disk group.

From SMU guide:

Unlike tiering, where a single copy of specific blocks of data resides in either spinning disks or SSDs, the Read Flash
Cache (RFC) feature uses one SSD read-cache disk group per pool as a read cache for “hot” pages only. Each read-cache
disk group consists of one or two SSDs with a maximum usable capacity of 4TB. A separate copy of the data is also kept
in spinning disks. Read-cache contents are lost when a controller restart or failover occurs.

Disabling cache flush is not a configurable option in MSA.

The cache-flush option listed under show cache-parameters command output is not a user configurable option.

[SUBHAJIT KHANBARMAN_1 · Accepted solution]

I understand your requirement is not to clean the data from SSD cache when power lost.

Please note that SSD cache actually nothing but Read Flash Cache (RFC) which is related to read operation which means data is already in backend drives. When host requests some data as read operation that time if the data not in read cache that time Controller fetch that data from backend drive to RFC. Now in this situation if Controller lost power you will lost the data  temporarily which you had fetched in RFC but actually the data intact in backend drives. When power regain that time it will again fetch same data for next read operation if same data requested by Hosts. This is absolutely fine and not require any change.

Now coming to your other query "

when executing the command: show cache-parameters is displayed as Cache Flush: Enabled. In the CLI reference guide, I did not find this option disabled.

"

Please note that this cache is actually called CompactFlash cache which comes into picture when Controller lost power and write data which is not yet written in backend drives will be kept safe.  Controller Cache memory is flushed to CompactFlash in the case of a controller failure or power loss. During the write to CompactFlash process, only the components needed to write the cache to the CompactFlash are powered by the supercapacitor. This process typically takes 60 seconds per 1Gb of cache. After the cache is copied to CompactFlash, the remaining power left in the supercapacitor is used to refresh the cache memory. While the cache is being maintained by the supercapacitor, the Cache Status LED flashes at a rate of 1/10 second on and 9/10 second off.

Now if you disable this option then your system will always work in Write-Through mode for which you will face performance issue.

Hope this helps!
Regards
Subhajit

I am an HPE employee

If you feel this was helpful please click the KUDOS! thumb below!

***********************************************************************************

[Mike91]

thanks for the clarification